AI: Who Still Uses Permissions?
TL;DR
A Hacker News user asks whether anyone still works with permission prompts in AI coding agents. Frustrated by harnesses that keep asking for tool approval and fail to remember granted permissions, the poster has run Claude Code with --dangerously-skip-permissions and Codex with an automatic approvals reviewer for about a year. The thread asks whether broken permission management is pushing developers to switch safety prompts off entirely.
Nauti's Take
The thread points to real potential for better tooling: agents that store permissions persistently and granularly would save developers a lot of clicking. The risk lies in the workaround, since an agent without prompts can delete files or read secrets unnoticed.
A sensible middle path is to run agents in a container or sandbox and allow automatic approvals only there.
Sources